Date: Mon, 26 Apr 1999 07:41:17 +0300 (EEST) From: Juha Nurmela <junki@qn-lpr2-98.quicknet.inet.fi> To: multimedia@FreeBSD.ORG Subject: Re: brooktree driver problems?? Message-ID: <Pine.BSF.4.05.9904260715220.22043-100000@pena.oh5nxo.ampr.org>
next in thread | raw e-mail | index | archive | help
There are sizes which don't generate additional errors, being careful to not to ask for odd dimensions (in &= 1 sense) I never saw more than 75 ints/sec with the original driver (from VSYNC twice and the final GEN_IRQ), after some bastardizing now only 25 ints/sec (GEN_IRQ only). Quality of the broadcast/camera might have an impact here... IMHO the risc code might be changed a bit to just programatically generate the wanted interrupts (with risc-status) and mask away vsync ints. Input errors (loss of signal etc) should be taken too. The complicated determination of correct field/frame in bktr_intr would then be unnecessary as the risc code would not proceed into generating the correct risc-status irq without good signal flowing. I got so confused about the even/odd things. Does VRE _follow_ even field ? Which field produces the first visible scanline ? Circles in still pictures look correct but moving objects generate an effect which led me to think the fields are from differing frames (maybe this is just the way television works ?) Just try to capture interlaced picture from the fast-moving business figures on CNBC :) Can we get vbi data without the corresponding field of pixel data ? Juha To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-multimedia" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.4.05.9904260715220.22043-100000>